This repository is a ROS 2 package that can receive data from various motion capture systems:
For most systems, three different tracking modes are available:
The data is directly published via tf2 and a /poses
topic that supports different QoS settings.
This package was originally developed for Crazyswarm to track up to 50 flying robots.
To build from source, clone the latest version from this repository into your ROS 2 workspace and compile the package using
cd ros_ws/src
git clone --recurse-submodules https://github.com/IMRCLab/motion_capture_tracking
cd ../
colcon build
ros2 launch motion_capture_tracking launch.py
The various options can be configured in config/cfg.yaml
.
There are two possible backends.
Make sure that you have the following settings in Motive:
menu Edit/Settings/Streaming:
We recommend that you first try "optitrack" and switch to "optitrack_closed_source" if you encounter any issues.
Since the SDK is not publically available, adding the SDK and building from source is required.
motion_capture_tracking/deps/libmotioncapture/deps/nokov_sdk
(such that there is for example the file motion_capture_tracking/deps/libmotioncapture/deps/nokov_sdk/lib/libSeekerSDKClient.so
)motion_capture_tracking/CMakeLists.txt
change set(LIBMOTIONCAPTURE_ENABLE_NOKOV OFF)
to set(LIBMOTIONCAPTURE_ENABLE_NOKOV ON)
colcon build
The ROS package is a wrapper around libmotioncapture and librigidbodytracker. The former is a C++ library that provides a unified interface over different motion capture SDKs to gather pose informations of rigid bodies and/or pointclouds of unlabeled markers. The latter is a C++ library that takes the following inputs: i) a first-order dynamics model, ii) initial poses of rigid bodies, and iii) at every frame a point cloud. It outputs for every frame the best estimate of the robots' poses.
Some more information on the rigid body pose tracking is available in
@inproceedings{crazyswarm,
author = {James A. Preiss* and
Wolfgang H\"onig* and
Gaurav S. Sukhatme and
Nora Ayanian},
title = {Crazyswarm: {A} large nano-quadcopter swarm},
booktitle = {{IEEE} International Conference on Robotics and Automation ({ICRA})},
pages = {3299--3304},
publisher = {{IEEE}},
year = {2017},
url = {https://doi.org/10.1109/ICRA.2017.7989376},
doi = {10.1109/ICRA.2017.7989376},
note = {Software available at \url{https://github.com/USC-ACTLab/crazyswarm}},
}
The unlabeled marker tracking is using an optimal assignment with a min-cost max-flow formulation for each frame.
These are current alternatives if no custom tracking is needed: